There is a problem for both pools: too often (half second to 1 second), pool re-send the same block header data to miners, which causes too many repeat calculation on block header data. Data is not steady for continue mining. It is not good for small CPU, and it produces too much internet traffic, I believe.
It has next to zero effect.
Even an RPi CPU can do millions of hashes a second.
Merkle slice work generation takes about 10 hashes.
However, how often?
My pool sends new work every 30 seconds with clean=false
Which is on purpose, by design and necessary for bitcoin.
and of course sends work on a network block change = about 144 times a day
... for the rather obvious reason
However, this ~144 times a day sends a clean=true work item coz working on stale blocks is a
REALLY bad idea,
followed by a clean=false work item.
I can only guess you mean this extra clean=false work item about 144 times a day?
Edit: and total network is about ... 15Mbytes a day. That one extra work item is about 0.2Mbytes a day
Yikes! That's called too much?